1. Welcome to Mu-43.com—a friendly Micro 4/3 camera & photography discussion forum!

    If you are thinking of buying a camera or need help with your photos, you will find our forum members full of advice! Click here to join for free!


Discussion in 'Astrophotography' started by Darin Berdinka, Jan 10, 2019.

  1. I've just recently gotten back into photography after a number of years where I slowly devolved to an iphone shooter. While the weather was clear in December I started trying to do some astrophotography with my e-m1 ii. The simple stuff, single frame or stacked frame with exposures in the 10-20 second range.

    I read what I could find online, did the iso invariance test and concluded that in low light noise reached a minimum at 800 or 1600 iso with my camera. 1600 seemed to have slightly less noise and 800 seemed to retain detail slightly better, but really it was a wash as far as my eyes could tell. I decided to shoot as iso 800 under the logic that I would minimize overexposure and cooresponding CA around really bright stars. Seems to work, I didn't get an amazing shots, but I definitely got some good ones.

    Getting to the point.....

    Looking at other peoples m43 images I see a wide range of iso being used from 800 all the way up to 12,800. If my understanding of how iso works is correct there's no extra signal in higher iso shots just amplification of the same signal. This amplification can be employed in pp as well with, I reason, better control over highlights as so forth. Am I missing something?

    So what iso do you shoot at and why?
  2. You have missed something. The gain stages are all the way back in hardware at the sampling stage of the pipeline prior to RAW capture. This is not the same as applying gain in software at the end of the pipeline. Once you've lost signal to noise ratio at the start and then bake it in the RAW you can't get it back later.

    Now there is this thing called ISO-invariance which is as you've described to some extent. For a certain range of ISOs on specific sensors that exhibit this phenomenon, the gain is applied somewhere later in the pipeline so pushing in post has basically the same effect (only for the range of ISOs where invariance is exhibited though).

    Even with ISO-invariance, you have to be careful since you may end up burying the shadows too far into the noise floor or even clipping it. Such shadows are not recoverable.
    • Agree Agree x 1
  3. ralf-11

    ralf-11 Mu-43 All-Pro

    Jan 16, 2017
    DR decreases as ISO increases ----> use ISO 200 if possible; OTOH you may have to exposure stack anyway if the sky is in your Landscape pics

    ISO 3200 is about the limit for good IQ pics, so I set the upper limit there
  4. Bushboy

    Bushboy Mu-43 Top Veteran

    Apr 22, 2018
    Auto iso up to 3200.
    I let the camera decide.
    I have enough to think about.
  5. pellicle

    pellicle Mu-43 Hall of Famer Subscribing Member

    Feb 10, 2010
    Killarney, OzTrailEYa

    I'm curious how you concluded this ...
    because considering the above you may have concluded incorrectly looking at JPG's and having fine details squished out.

    The best way to proceed is to use RAW and take a single shot at each of the desired settings (shutter speed and ISO) then look at that developed in something Plane Jane like DCRAW (because that's exactly the engine that stacker programs will likely be based on, its pretty much the norm under the hood).

    Stacking will reduce the noise (meaning give you a better signal to noise ratio) and you do need to toss in a few dark frames too.

    Myself I process in Deep Sky Stacker only to combine the images, then put that TIFF through Photoshop to trim up the dark and light points where I have better control (remember to adjust level dark and light points by holding down Alt first then moving the dark or light slider with the mouse keeping Alt held down).
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.